Skip to content. Skip to navigation
McGill Home SOCS Home
Personal tools
You are here: Home Announcements and Events Colloquium Archive

Colloquium Home
Fall 2013 Schedule
Winter 2014 Schedule
How to Suggest a Speaker


( Fall 2012 )
Speaker and Abstract
2012/09/17 Speaker: Srini Devadas
Affiliation: MIT
Title: EM2, Angstrom and Ascend: Processor Building at MIT
Abstract: We describe three processors in various stages of design and implementation at MIT.
The Execution Migration Machine (EM2) a computation-migration-based multicore architecture that provides speedy access to on-chip distributed cache data by either migrating execution or via remote memory operations. A 121-core chip is slated for tapeout in January 2013 in 45nm technology.
Angstrom is a proposed design of a 1000-core processor that includes self-awareness in the processor, memory and network subsystems to address the energy wall problem in multicores. We have recently fabricated a "tile" of Angstrom that includes self-aware processor features.
Ascend -- Architecture for Secure Computation on Encrypted Data -- is a new processor architecture that guarantees security of encrypted data even when untrusted application and/or system software executes on the data. Ascend never exposes decrypted data to the server and also hides memory access patterns through obfuscated instruction execution. We have performed detailed simulations of an Ascend processor design to evaluate overheads of encrypted computation.
Biography of Speaker:

Srini Devadas is a Professor of Electrical Engineering and Computer Science at the Massachusetts Institute of Technology (MIT), and has been on the faculty of MIT since 1988. He served as the Associate Head with responsibility for Computer Science from 2005-2011. Devadas has worked in the areas of Computer-Aided Design, testing, formal verification, compilers for embedded processors, computer architecture, computer security, and computational biology and has co-authored numerous papers and books in these areas. Devadas was elected a Fellow of the IEEE in 1998.

2012/09/21 Speaker: Arun Somani
Affiliation: Iowa State University
Area: Networking
Title: Simple Problems Complex Solutions: Issues in (Optical) Networking
Abstract: Today’s Internet is being used to support an array of new services and applications. Many of these applications pose a variety of performance, routing, and reliability requirements. Individual applications have different bandwidth requirements, varying from very low to very high, that needs to be supported over the high capacity wavelength channels in a WDM optical network. We consider network architectures supported by an appropriate control plane that is capable of providing efficient scheduling, grooming, multicasting, and protection mechanism to support the demanding requirements of emerging applications. We consider several issues and possible solutions. Our abstraction captures the basic properties of links and nodes, but is not excessively burdensome by requiring the maintenance of excessive state information. The concepts are transparent, scalable, reliable, and simple, and can support dynamic traffic. The talk will include a short introduction to optical networking for a wider audience. Biography of Speaker:

Arun K. Somani is currently Anson Marston Distinguished Professor of Electrical and Computer Engineering at Iowa State University, Ames, IA. He earned his MSEE and PhD degrees in electrical engineering from the McGill University, Montreal, Canada, in 1983 and 1985, respectively. He has worked as Scientific Officer for Govt. of India, New Delhi from 1974 to 1982 and as a faculty member at the University of Washington, Seattle, WA from 1985 to 1997. Professor Somani's research interests are in the area of computer system design and architecture, dependable computing and networking, WDM-based optical networking, and reconfigurable and parallel computer systems and use of information technology for infrastructure applications. He has ~300 technical papers, several book chapters, one book, and has supervised more than 60 MS and more than 25 PhD students. He has served on several program committees in various capacities, IEEE distinguished visitor and IEEE distinguished tutorial speaker. He has delivered several key note speeches, tutorials and distinguished and invited talks all over the world. In 1999, he was elected a Fellow of IEEE for his contributions to “theory and applications of computer networks.” He has been awarded a Distinguished Scientist member grade of ACM in 2006.

2012/09/28 Speaker: Yi-Min Wang
Affiliation: Microsoft Research-Redmond (MSRR)
Title: Having Fun, Taking Risks, and Changing the World at Microsoft Bldg. 99
Abstract: Some say Pleasure, Engagement, and Meaning are the three components of happiness. Research minds want to pursue them all. They want the freedom to choose to work on topics that they can have the most fun with. They want to take risks by constantly challenging themselves intellectually with full engagement or even obsession. They want to make a big difference in the world by impacting the lives of millions or even billions of people. In this talk, I will share the stories of some very happy researchers at Microsoft Bldg. 99 and how we strive to create an environment to maximize research happiness. Biography of Speaker:

Dr. Yi-Min Wang is currently Deputy Managing Director of Microsoft Research-Redmond, co-managing 300+ researchers and engineers across a wide range of EE and CS disciplines. Prior to that, he was Director of ISRC (Internet Services Research Center) where he led a team focusing on developing web-scale data technologies and contributing significantly to the quality of Microsoft’s search engine. Yi-Min was elected to IEEE Fellow for his contributions in the areas of dependable computing and web security. In 2005, he invented Strider HoneyMonkey - the first automated system to patrol the Web and hunt for malicious websites that exploit zero-day vulnerabilities. The HoneyMonkey technology has become the de facto standard for both the security industry and the search engine industry. In 2007, he invented Strider Search Ranger – the first search-spam detection system based on dynamic crawling and traffic analysis. The work was featured on New York Times and has had an industry-wide impact on wiping out search-spam. Yi-Min received a bachelor of science degree in electrical engineering from National Taiwan University in 1986, a Master’s degree in electrical engineering (with a thesis on spectral estimation) in 1990, and a doctorate in computer engineering (with a thesis on fault tolerance) from University of Illinois at Urbana-Champaign in 1993.

2012/10/05 Speaker: Magdalena Balazinska
Affiliation: University of Washington Seattle
Area: Databases
Title: Big Data Management Promises and Challenges
Abstract: Our society is generating data at an unprecedented scale and rate. The need to manage this data onslaught is growing in both industry and science. The ability to manage "Big Data" holds the promise to deliver novel services and accelerate discovery. Today's data management systems, however, have several limitations. In this talk, we discuss two types of limitations and examples of techniques aimed at addressing each one of them. First, we discuss the problem of efficient parallel data processing focusing on failure-handling, which is critical at massive scale. Second, we discuss the problem of efficient, out-of-the-box processing of user-defined operations, which commonly exhibit load imbalances. We conclude with a brief discussion of additional challenges related to Big Data management today. Biography of Speaker:

Magdalena Balazinska is an Associate Professor in the department of Computer Science and Engineering at the University of Washington. Magdalena's research interests are broadly in the fields of databases and distributed systems. Her current research focuses on big data management, sensor and scientific data management, and cloud computing. Magdalena holds a Ph.D. from the Massachusetts Institute of Technology (2006). She is a Microsoft Research New Faculty Fellow (2007), received an NSF CAREER Award (2009), a 10-year most influential paper award (2010), an HP Labs Research Innovation Award (2009 and 2010), a Rogel Faculty Support Award (2006), a Microsoft Research Graduate Fellowship (2003-2005), and multiple best-paper awards.

2012/10/12 Speaker: Karan Singh
Affiliation: University of Toronto
Area: Computer Graphics
Title: Art and Perception driven Interactive Modeling
Abstract: Sketch, sculpt and touch interfaces have often been touted as "natural" approaches to interactive design. While these metaphors are indeed a promising medium of visual communication, there are a number of inherent limitations in the motor control of the human hand, drawing or gesturing skill, perception and the ambiguities of inference, that make the leap from 2D input to 3D shape modeling a challenging task. In this talk I will present recent research and open challenges in the perception of shape from sketch/sculpt style input and various systems (,, that facilitate the leap from 2D input to 3D models despite these limitations. Biography of Speaker:

Karan Singh is an an Associate Professor in Computer Science at the University of Toronto. He holds a BTech. from IIT Madras, MS, PhD from the Ohio State University. His research interests lie in artist driven interactive graphics, spanning character animation, anatomic modeling, geometric shape design and sketch-based interfaces. He has been a technical lead the Oscar winning animation system Maya. He co-directs the graphics and human computer interaction lab, DGP and was the R&D Director for the 2005 Oscar winning Animation film “Ryan”.

2012/10/19 Speaker: Max Schaefer
Affiliation: IBM Research
Area: Programming Languages
Title: Efficient Construction of Approximate Call Graphs for JavaScript IDE Services
Abstract: The rapid rise of JavaScript as one of the most popular programming languages of the present day has led to a demand for sophisticated IDE support similar to what is available for Java or C#. However, advanced tooling is hampered by the dynamic nature of the language, which makes any form of static analysis very difficult. We single out efficient call graph construction as a key problem to be solved in order to improve development tools for JavaScript. To address this problem, we present a scalable field-based flow analysis for constructing call graphs. Our evaluation on large real-world programs shows that the analysis, while in principle unsound, produces highly accurate call graphs in practice. Previous analyses do not scale to these programs, but our analysis handles them in a matter of seconds, thus proving its suitability for use in an interactive setting. Biography of Speaker:

Max Schaefer is currently a post-doctoral researcher at the IBM T.J. Watson Research Center, working mostly on program analysis and transformation, particularly for dynamic programming languages. Before coming to IBM, he obtained his DPhil from the University of Oxford, graduating with a thesis on specification, implementation and verification of refactorings.

2012/10/26 Speaker: Tim Roughgarden
Affiliation: Stanford
Area: Algorithms
Title: Porting the Computer Science Toolbox to Game Theory and Economics
Abstract: Theoretical computer science has brought new ideas and techniques to game and economic theory. A primary signature of the computer science approach is {em approximation} --- the idea of building credibility for a proposed solution by proving that its performance is always within a small factor of an ideal (and typically unimplementable) solution. We explain two of our recent contributions in this area, one motivated by networks and one by auctions. We first discuss the "price of anarchy": how well does decentralized (or "selfish") behavior approximates centralized optimization? This concept has been analyzed in many applications, including network routing, resource allocation, network formation, health care, and even models of basketball. We highlight a new theory of robust price of anarchy bounds, which apply even to systems that are not in equilibrium. Second, we consider auction design: for example, what selling procedure should be used to maximize the revenue of a seller? On the analysis side, we highlight a new framework that explicitly connects average-case (i.e., Bayesian) analysis, the dominant paradigm in economics, with the worst-case analysis approach common in computer science. On the design side, we provide a distribution-independent auction that performs, for a wide class of input distributions, almost as well as the distribution-specific optimal auction. Biography of Speaker:

Tim Roughgarden received his Ph.D. from Cornell University in 2002 and joined the Stanford CS department in 2004, where he is currently an associate professor. His research interests are in theoretical computer science, especially its interfaces with game theory and networks. He wrote the book "Selfish Routing and the Price of Anarchy" (MIT Press, 2005) and co-edited the book "Algorithmic Game Theory", with Nisan, Tardos, and Vazirani (Cambridge, 2007). His awards include the 2002 ACM Doctoral Dissertation Award (Honorable Mention), the 2003 Tucker Prize, a 2007 PECASE Award, the 2008 Shapley Lectureship of the Game Theory Society, the 2009 ACM Grace Murray Hopper Award, and the 2012 EATCS-SIGACT Godel Prize. He recently developed a free online course on the design and analysis of algorithms, enrolling tens of thousands of students.

2012/11/02 Speaker: Michiel van de Panne
Affiliation: University of British Columbia
Area: Computer Animation
Title: Learning to Locomote
Abstract: The ability to move and act in the physical world is fundamental to humans and animals Yet our best models for simulating their motion or controlling legged robots are still a pale imitation of the agility and grace of motions seen in nature. Is the missing ingredient a matter of computation, mechanics, data, sensing, learning, or something else? I will describe our own work towards answering these questions in a series of recent projects that demonstrate a large variety of skills for physics-based simulations of humans and canines. Our solutions argue for a number of relevant abstractions and the incremental development of motion skills. We further argue that physics-based animation provides a fertile ground for modeling a growing range of motion skills. It provides new tools and approaches towards the control problems that it shares with biomechanics and robotics, and is replete with challenges for planning and learning algorithms. The future is agile and dexterous. Biography of Speaker:

Michiel van de Panne's research interests are in physics-based animation and simulation of characters, computer graphics, motion planning and control, robotics, sketch-based modeling, and applications of machine learning to computer graphics and animation. He recently completed a ten year term term as a Tier 2 Canada Research Chair in Computer Graphics and Animation. In 2002 he co-founded the ACM/Eurographics Symposium on Computer Animation (SCA), the leading forum dedicated to computer animation research. He served as an Associate Editor of ACM Transactions on Graphics during 2005-2008. He has co-chaired EG CAS 1997, SCA 2002, GI 2005, SBIM 2007, and SCA 2011. He serves on the program committees of ACM SIGGRAPH, Eurographics, ACM/EG SCA, ACM I3D, Graphics Interface, and NPAR. The work he did with his M.Sc. student Ivan Neulander helped form the basis of the Rhythm & Hues hair rendering pipeline for The Chronicles of Narnia and other films.

2012/11/09 Speaker: Jon Sporring
Affiliation: Department of Computer Science at the University of Copenhagen
Area: Computer Graphics
Title: Spatial and Temporal Photon Differentials
Abstract: We consider ray bundles emanating from a source such as a camera or light source, and we derive the full spatial and temporal structure to first order of the intersection of ray bundles with scene geometry where scene geometry is given as any implicit function. Further, we present the full details of two often used geometrical representations. The first order structure may be used as the linear approximation of the change of photon shape as the camera, objects, and light source change as function of space and time. Our work generalizes previous work on ray differentials and photon differentials both in space and time. Biography of Speaker:

Jon Sporring received his Master and Ph.D. degree from the Department of Computer Science, University of Copenhagen, Denmark in 1995 and 1998, respectively. Part of his Ph.D. program was carried out at IBM Research Center, Almaden, California, USA. Following his Ph.D, he worked as a visiting researcher at the Computer Vision and Robotics Lab at Foundation for Research & Technology - Hellas, Greece, and as assistant research professor at 3D-Lab, School of Dentistry, University of Copenhagen. Since 2003 he has been employed as associate professor at the Department of Computer Science, University of Copenhagen. From 2008-2009 he was part-time Senior Researcher at Nordic Bioscience a/s. In the period 2007-2012 he was Vice-Chair for Research at Department of Computer Science, and since 2012 he is a visiting professor at School of Computer Science, McGill University, Montreal, Canada. His main topics of research are image processing and computer graphics.

2012/11/16 Speaker: SOCS Professors
Affiliation: School of Computer Science, McGill University
Title: SOCS Faculty Research Overviews
Abstract: Prof. Wenbo He will be hosting a special colloquium featuring short presentations from a number of faculty members in the school of computer science. This is your opportunity to get a quick overview of the research that is happening in the department. Professors will talk about their research programs and recently published work, but more importantly they will also speak about current projects and future opportunities for new students. Come hear about new research problems that are not yet described on any public web page, and then discuss with students and faculty over snacks and drinks at the reception that follows. There will be research overviews given by the following professors: Wenbo He, Mike Langer, Greg Dudek, Derek Ruths, Hamed Hatami. Biography of Speaker:

2012/11/30 Speaker: Monty Newborn
Affiliation: McGill University
Title: An Overview of Computer Chess: History and Lessons
Abstract: This talk will survey the history of computer chess from Charles Babbage to IBM's Deep Blue and ending with the current year. It will attempt to cover some of the general artificial intelligence lessons that have been learned as a result of years of research and development. Once up a time the top chess players laughed at the play of computers. Then in the 1980s and 1990s, they came to watch how well they were doing. Today, they are learning from computers. While chess players are learning to play chess by competing with computers, computer scientists are using the knowledge gained when designing other AI systems. This talk will assume the audience is not composed of chess experts! The lessons are related to: Intuition, programming languages, computer speed, brute-force search (versus selective search), hash tables, iteratively deepening search, and endgame play. Pre Deep Blue history includes the ideas of: Babbage, Turing, Shannon, McCarthy, Slate, Thompson (Belle), and Bob Hyatt (Cray Blitz). Deep Blue’s two matches: First match with Kasparov in 1996 – Kasparov won 4-2, Second match with Kasparov in 1997 – Deep Blue won 3.5 – 2.5. Post Deep Blue history: Three major man-machine matches – man unable to win. Increasingly stronger programs—Fritz, Shredder, Junior, Hydra, Zappa, Rybka, Houdini. Biography of Speaker:

Newborn's research has focused on search problems in artificial intelligence where two areas have been of particular interest: chess-playing programs and automated theorem-proving programs. He has published eight books on these subjects and a number of research papers as well. His most recent book published by Springer is entitled Beyond Deep Blue. He served as chairman of the ACM Computer Chess Committee from 1981 until 1997. In that capacity he organized the first Kasparov versus Deep Blue match (known as the ACM Chess Challenge) in 1996. The following year he served as head of the officials at the second Kasparov versus Deep Blue match won by Deep Blue. Throughout the 1970s and 1980s, his chess program Ostrich competed in five world championships, coming close to winning the first one in 1974. Newborn also developed two automated theorem-proving programs, Theo and Octopus. Octopus is a multiprocessor version of Theo and has run on more than 100 PCs during past ATP competitions.

2012/12/07 Speaker: Charles Consel
Affiliation: University of Bordeaux I
Title: Design-driven Development of Dependable Applications: A Case Study in Avionics
Abstract: Making an application dependable demands that its functional and non-functional requirements be stringently fulfilled throughout its development process. In this context, a design-driven development approach has the key advantage of enabling requirements to be traced from their high-level design forms to the resulting executable artifact. However, because such approaches are mostly general purpose, they provide little design guidance, if any. This situation makes unpredictable the coherence and the conformance of an application with respect to its requirements. To address this situation, we propose an approach that leverages a design-driven development process dedicated to a specific paradigm. This approach guides the verification of the coherence and conformance of an application throughout its development. We demonstrate the benefits of our approach by applying it to a realistic case study in the avionics domain. Biography of Speaker:

Charles Consel is a professor of Computer Science at University of Bordeaux I. He served on the faculty of Yale University, Oregon Graduate Institute and the University of Rennes. He leads the Phoenix group at INRIA. He has been designing and implementing Domain-Specific Languages (DSLs) for a variety of areas including device drivers, programmable routers, stream processing, and telephony services. These DSLs have been validated with real-sized applications and showed measureable benefits compared to applications written in general-purpose languages. His research contributions cover programming languages, software engineering, operating systems, pervasive computing, and assisted living.